Yes— but with one important caveat: I would describe Innodata more as a “small Palantir in the data/AI backend” rather than a mini-Palantir in terms of its actual business model.
In my view, the comparison is particularly interesting precisely because of Innodata’s recent developments:
The key point
is that Innodata is one level below Palantir .
Palantir, for example, takes corporate or government data and builds operational decision-making and AI systems from it.
Innodata, on the other hand, ensures that the data, training datasets, evaluations, and AI workflows actually meet the necessary quality standards.
This becomes particularly clear through a remarkable connection:
Innodata is itself a supplier to Palantir.
Palantir has selected Innodata for data engineering, annotation, and multimodal data processing—including computer vision, video, image, document, and sensor data.
I find this strategically very interesting.
And Innodata is continuing to evolve
The company is no longer just a traditional “data processor.”
In Q2 2026:
- Revenue +58%
- Adjusted EBITDA +92%
- Gross margin 49%
- Largest customer now only 37% of revenue, down from 56% in Q1
- New major Frontier AI customer
- Activities at Agentic AI / Reinforcement Learning
- AI cyber training
- Data for Robotics / Physical AI
- Expansion of evaluation and observability for AI agents.
Innodata – Innodata meldet Rekordergebnisse für das zweite Quartal 2026
Even more interesting: In Q1, a Evaluation & Observability Platform was introduced in Q1, designed to function as a kind of control layer for agent-based AI systems.
So I would put it this way:
Palantir = AI operating system for businesses and governments
Innodata = data/AI infrastructure and quality layer upon which such systems can be built
And that’s exactly where I see the exciting investment story at Innodata.
Because as AI agents, physical AI, robotics, defense AI, and government AI grow rapidly, the demand increases not only for computing power and models—but also for high-quality data, evaluation, security, and continuous training.
That is precisely the layer on which Innodata is increasingly focusing. The company now expects at least 40% revenue growth.
For the tenbagger thesis
I would therefore not simply buy Innodata as a “mini-Palantir”, but rather frame the thesis a bit differently:
“Innodata is a potential ‘picks-and-shovels’ beneficiary of the Palantir/AI revolution.”
And to me, that’s actually the more interesting thesis.
After all, Innodata doesn’t have to become the dominant AI platform itself. It’s enough if more and more AI companies and government organizations require high-quality data, agent training, evaluation, and secure AI workflows .
The fact that Palantir is already an Innodata customermakes this theory particularly plausible.
